G1 Robot Operation Document

Teleoperation and G1 handling guide.

---

Table of Contents

1. G1 Humanoid Robot Guide
2. Installation
3. Usage
4. Network Configuration
5. Safety
6. Troubleshooting
7. Inspire Hands Gen 4

---

G1 Humanoid Robot Guide

This guide covers the essential setup and operation of the Unitree G1 Humanoid Robot.

Prerequisites

Required Skills

• Linux: Basic command line skills.
• ROS2: Familiarity with ROS 2 concepts (nodes, topics).
• Programming: Python or C++ for custom development.

PC Minimum Requirements

• RAM: 8 GB+
• Storage: 50 GB free
• GPU: no GPU requirement unless adding some 3rd party policies
• OS: Ubuntu 22.04 LTS

Remote & App Control

For assistance with the remote controller or mobile app, contact Soofiyan.

---

Installation

Follow these steps to set up the control environment and necessary software packages.

1.1. Inverse Kinematics dependencies

First, create and activate a new Conda environment. Then, install the required packages.

$ conda create -n tv python=3.10
$ conda activate unitree
(unitree) $ conda install pinocchio -c conda-forge
(unitree) $ pip install meshcat
(unitree) $ pip install casadi

1.2. Unitree SDK

Next, clone and install the unitree_sdk2_python library.

# Install unitree_sdk2_python.
(unitree) $ git clone [https://github.com/unitreerobotics/unitree_sdk2_python.git](https://github.com/unitreerobotics/unitree_sdk2_python.git)
(unitree) $ cd unitree_sdk2_python
(unitree) $ pip install -e .

1.3. Teleoperation Dependencies Setup

Finally, clone the g1 teleoperation repository and install its specific requirements.

(unitree) $ cd ~
(unitree) $ git clone [email protected]:Soofiyan/g1_teleoperate.git
(unitree) $ cd ~/g1_teleoperate
(unitree) $ pip install -r requirements.txt
(unitree) $ pip install empy==3.3.4
(unitree) $ pip uninstall typing

1.4. (Optional) ROS2 install

Go through the README.md file of this repository: unitree_ros2.

---

Usage

How to run the robot ik on meshcat

# Go to the g1_teleoperate root folder
(unitree) $ cd g1_teleoperate/teleop/robot_control
(unitree) $ python robot_arm_ik_og.py

---

Network Configuration

This section explains how to establish a direct network connection between your computer and the robot.

Configuration Steps

1. Connect a network cable from the robot to your computer.
2. Enable USB Ethernet on your computer.
3. The robot's IP address is fixed at 192.168.123.161. You must set your computer's USB Ethernet IP to the same subnet. Set your computer's IP to 192.168.123.222.

Testing the Connection

To verify the connection, ping the robot from your computer's terminal:

ping 192.168.123.161

A successful reply indicates the connection is working.

Identify Your Network Device

You need to find the network card name assigned to this connection. Use the ifconfig command.

In the example below, the network card name for the IP 192.168.123.222 is enxf8e43b808e06. Note this name, as it's required to run the control examples.

Safety

(This section is under development.)

---

Troubleshooting

(This section is under development.)

Inspire Hands Gen 4

Setup and Installation

Follow these steps to set up the necessary software for the Inspire Hand.

# 1. Clone the repository
git clone https://github.com/Soofiyan/inspire_hand_ws.git
cd inspire_hand_ws

# 2. Install the Unitree SDK
cd unitree_sdk2_python
pip install -e .

# 3. Install the Inspire Hand SDK
cd ../inspire_hand_sdk
pip install -e .

# 4. Install required Python libraries
pip install pynput
pip install pymodbus==3.6.9

---

Read Sensor Data

Use the data_handler_worker callback function from the sub_traget_LR.py code to read sensor data.

Control Modes

The Inspire Hand SDK offers various control modes. You can select a mode by its corresponding ID number.

Mode ID	Binary	Description
0	0000	No Operation
1	0001	Angle Control
2	0010	Position Control
3	0011	Angle + Position
4	0100	Force Control
5	0101	Angle + Force
6	0110	Position + Force
7	0111	Angle + Position + Force
8	1000	Velocity Control
9	1001	Angle + Velocity
10	1010	Position + Velocity
11	1011	Angle + Position + Velocity
12	1100	Force + Velocity
13	1101	Angle + Force + Velocity
14	1110	Position + Force + Velocity
15	1111	Angle + Position + Force + Velocity

---

© ARClab. All Rights Reserved.